Arpit Agarwal created HDDS-1429:
-----------------------------------
Summary: Avoid using common fork join pool
Key: HDDS-1429
URL: https://issues.apache.org/jira/browse/HDDS-1429
Project: Hadoop Distributed Data Store
Issue Type: Improvement
Reporter: Arpit Agarwal
Assignee: Arpit Agarwal
After enabling thread context in Ozone log messages, we see some Ratis
operations being executed in common fork join pool. E.g.
{code}
2019-04-11 14:25:54,583 ForkJoinPool.commonPool-worker-3 INFO
storage.RaftLogWorker (RaftLogWorker.java:rollLogSegment(303)) -
3e59bcbc-e6f0-4681-b8a6-76e163e9ff19-RaftLogWorker: Rolling segment log-0_70 to
index:70
{code}
and
{code}
2019-04-11 14:25:56,715 ForkJoinPool.commonPool-worker-1 INFO
client.GrpcClientProtocolService
(GrpcClientProtocolService.java:lambda$processClientRequest$0(264)) - Failed
RaftClientRequest:client-B5832CAE4B89->8a182bb4-96a8-42e8-a7da-549c9663fc30@group-7136CD304607,
cid=333, seq=0, Watch-ALL_COMMITTED(79), Message:<EMPTY>,
reply=RaftClientReply:client-B5832CAE4B89->8a182bb4-96a8-42e8-a7da-549c9663fc30@group-7136CD304607,
cid=333, FAILED org.apache.ratis.protocol.NotLeaderException: Server
8a182bb4-96a8-42e8-a7da-549c9663fc30 is not the leader
(beff1b4d-05f9-4f7d-a6f2-1405b0950e8c:10.22.8.149:57253). Request must be sent
to leader., logIndex=0, commits[8a182bb4-96a8-42e8-a7da-549c9663fc30:c129,
48a1cbdc-86a2-40d3-9831-13e044923ee4:c72,
beff1b4d-05f9-4f7d-a6f2-1405b0950e8c:c129]
{code}
It's better to use a dedicated ExecutorService or ForkJoinPool instead.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: hdfs-dev-unsubscr...@hadoop.apache.org
For additional commands, e-mail: hdfs-dev-h...@hadoop.apache.org
